Fazl Barez

Precise In-Parameter Concept Erasure in Large Language Models

arXiv 2025

Sycophancy to Subterfuge: Investigating Reward-Tampering in Large Language Models

arXiv 2024

Best-of-N Jailbreaking

arXiv 2024

Sleeper Agents: Training Deceptive LLMs that Persist Through Safety Training

arXiv 2024

Towards Interpreting Visual Information Processing in Vision-Language Models

arXiv 2024

Enhancing Neural Network Interpretability with Feature-Aligned Sparse Autoencoders

arXiv 2024

Quantifying Feature Space Universality Across Large Language Models via Sparse Autoencoders

arXiv 2024

Detecting Edit Failures In Large Language Models: An Improved Specificity Benchmark

arXiv 2023

Understanding Addition in Transformers

arXiv 2023

Interpreting Learned Feedback Patterns in Large Language Models

arXiv 2023

David Krueger

Ethan Perez

Philip Torr

Buck Shlegeris

Carson Denison

Clement Neo

David Duvenaud

Evan Hubinger

Jared Kaplan

co-founder / Chief Science Officer

Luke Marks